1 Evidence of a Latitudinal Gradient in the Age at Onset of Schizophrenia Andrew Shaner,* Geoffrey Miller† and Jim Mintz* *Department of Psychiatry and Biobehavioral Sciences, University of California, Los Angeles and the Veterans Affairs Greater Los Angeles Healthcare System, USA †Department of Psychology, University of New Mexico, USA 25 March 2007 In press for Schizophrenia Research Corresponding Author: Andrew Shaner, M.D. Deputy Chief of Psychiatry and Mental Health (116A) VA Greater Los Angeles Healthcare System 11301 Wilshire Blvd. Los Angeles, CA 90073 After publication correspondence and requests for materials should be addressed to A.S. (e-mail: Andrew.Shaner@med.va.gov), G.M. (e-mail: gfmiller@unm.edu) or J.M. (email: JMintz@ucla.edu). 2 Abstract: Variation in the age at onset of a multifactorial disease often reflects variation in cause. Here we show a linear latitudinal gradient in the mean age at onset of schizophrenia in 13 northern-hemisphere cities, ranging from 25 years old in Cali, Columbia (at 4 degrees north) to 35 years old in Moscow, Russia (at 56 degrees north). To our knowledge, this striking association has not been previously reported. We consider several explanations, including the effects of pathogen stress, natural selection, sexual selection, migration, life-history profiles, or some combination of these factors, and we propose a test of competing causal hypotheses. Keywords: Schizophrenia; latitudinal gradient; age at onset; evolution; sexual selection; fitness indicator. 3 1. Background Schizophrenia strikes about 1% of people worldwide, producing delusions, hallucinations, disorganized speech, bizarre behavior, and emotional blunting. Typically beginning in adolescence or early adulthood, it often leads to social isolation and severe lifelong disability. Though schizophrenia is moderately heritable (Cardno et al., 1999), no susceptibility genes of major effect have been clearly replicated (McDonald & Murphy, 2003). Its pathophysiology is poorly understood (Thaker & Carpenter, 2001), and treatment is far from satisfactory (Carpenter, 2005). Variation in age at onset is a potential clue to the causes of schizophrenia because factors that increase the risk of developing a multifactorial disease usually hasten onset, and earlier onset usually predicts more severe disease (Childs & Scriver, 1986). Both relationships are true for schizophrenia. For example, male sex, unmarried status, poor premorbid functioning, a family history of mental illness, birth complications and urban residence have all been associated with both earlier onset and more severe illness (Jablensky, 2000). However, the mechanisms by which these factors increase risk remain unknown despite extensive research. It is generally assumed that schizophrenia interferes with the social functioning required for marriage and that the effect would be stronger among those with earlier onset and more severe illness. This would account for the 15 to 73 percent reduction in marriage rates among those with schizophrenia (Haverkamp, Propping, & Hilger, 1982) as well as the associations among marital status, disease risk and age at onset. However, one analysis of large prospective data set has challenged those assumptions and suggested that marriage may protect against the illness (Jablensky & Cole, 1997). Examining the data in this report, we noticed another effect—those living in cities closer to the equator developed the illness earlier. Here we report that latitudinal gradient, discuss several possible mechanisms, and a method of determining which is responsible. 2. Methods The report by Jablensky and Cole (1997) listed age at onset of schizophrenia (AAO) in thirteen northern hemisphere cities. Subjects were 1431 individuals with recent-onset schizophrenia making their first treatment contact. AAO was defined as the age at first psychotic symptom and was covariance-adjusted for sex, marital status, premorbid personality traits and family history of mental illness. The original source of the data was the World Health Organization Collaborative Study of the Determinants of Outcome of Severe Mental Disorders conducted between 1976 and 1985 (Jablensky et al., 1992). In each city, investigators continuously monitored all helping agencies that were the likely first points of contact for individuals with schizophrenia. Participants were examined with a set of standardized instruments and reliability was tested and maintained by conducting joint interviews and by rating audio and video-taped interviews across sites. During the interview with the patient or informant, investigators explored and dated the occurrence of 25 early symptoms and signs of the disorder. The investigator’s estimate of the age at which each symptom first appeared had to be supported by a narrative summary of the early symptoms and signs, mode of onset, the progression of symptoms and any relevant circumstances. We obtained the latitude of each city from standard references. 4 3. Results The positive linear relationship between AAO and latitude is statistically significant and explains most of the variability of the data (t = 6.43, df = 11, p < .0001, R2 = 0.790). Although the six cities in the middle of the latitude distribution lie below the regression line, a test for nonlinearity was not significant, t = 2.02, df = 10, p = 0.071, nor was the Wald-Wolfowitz runs test (z = -0.71, p = .24). These results suggest that a linear relationship is sufficient to describe the relationship between AAO and distance from the equator. <Insert Figure 1 and legend about here> Earlier AAO closer to the equator, however, might be simply an artifact of higher mortality closer to the equator (Teriokhin, Budilova, Thomas, & Guegan, 2004), because a higher proportion of those destined to developed schizophrenia later in life would have died before disease onset. Median age of the entire population in each of the 10 countries containing the 13 cities ("The World Factbook," Accessed March, 2006) correlated significantly with both AAO (r = .69) and latitude (r = .79), and latitude correlated significantly with AAO (r = 0.89). Partial correlations indicate that the direct path from latitude to AAO is unaffected by controlling for median age (latitude and AAO controlling for median age: r = 0.79; latitude and median age controlling for AAO: r = 0.48; median age and AAO controlling for latitude: r = -0.07). We also created a correlation matrix (N = 10 countries) with the correlation structure described above and tested two path models using structural equation modeling. For the path in which latitude influences median age which then influences AAO, lack-of-fit chi-square = 8.34 (p = 0.004) and goodness of fit index = .71. For the path in which latitude influences both median age and AAO directly, lack-of-fit chi-square = 0.02 (p = 0.89) and goodness of fit index = .999. Thus, the model in which latitude influences both of the other variables directly fits very well, while the alternative fits poorly and is statistically rejected. These results suggest that the latitudinal gradient in AAO is not an artifact of a latitudinal gradient in the age structure of populations from which the subjects were drawn. 4. Discussion These results suggest that a factor that varies with latitude influences AAO. There are many possibilities for that factor and for the mechanism by which it influences AAO. The factor could be physical or biological, act on individuals or evolution and influence disease risk or the maturation of a trait related to schizophrenia. One possibility is that the factor is a known correlates of AAO. However, four are unrelated to the gradient, because in the report from which our data were drawn (Jablensky & Cole, 1997), AAO was statistically adjusted for sex, marital status, premorbid functioning, and family history of mental illness. AAO also has been associated with birth complications (Verdoux et al., 1997), urban residence (Marcelis, Navarro-Mateu, Murray, Selten, & van Os, 1998), AAO among affected relatives (DeLisi, 1992), and dopamine receptor genes (Dubertret, Gorwood, Gouya, Deybach, & Ades, 2001). It thus is possible that latitude influences AAO by altering birth complication 5 rates, an urban factor that influences risk, or the frequencies of genes influencing AAO (e.g., dopamine receptor genes). Also, several studies have reported an excess of winter and early spring births among those with schizophrenia which may be due to a latitudinal factor (Tochigi, Okazaki, Kato, & Sasaki, 2004). If this seasonal pattern is associated with later AAO, then the same factor may also be responsible for the gradient in AAO. A second possibility is that the factor is among the known correlates of latitude. Many features of the physical environment (e.g., photoperiod, climate) vary with latitude and might influence AAO. Moreover, these physical features produce latitudinal gradients in the biological environment that also could affect AAO. For example, the well-known pattern of greater species diversity toward the equator extends to microorganisms and other parasites (Nunn, Altizer, Sechrest, & Cunningham, 2005), including those that infect and kill humans (Gangestad & Buss, 1993; Guernier, Hochberg, & Guegan, 2004). This latitudinal gradient in pathogen diversity allows us to illustrate three general mechanisms by which an environmental factor that varies with latitude could influence AAO. 4.1 Direct Effects on Individuals First, the environmental factor could act directly on individuals to increase risk and decrease AAO. For example, prenatal infections are associated with greater risk for schizophrenia, perhaps because infection disrupts neurodevelopment (Brown et al., 2005). Thus, greater pathogen diversity closer to the equator could act directly on individuals to increase risk of schizophrenia and decrease AAO. 4.2 Natural Selection on Maturation Rate Second, the environmental factor could act indirectly by influencing the evolution of traits (and genes) related to AAO. For example, sexual maturation may be related to AAO because schizophrenia is rare before puberty and most cases begin between ages 15 and 26 (Stevens, 2002). Greater pathogen diversity closer to the equator may contribute to higher human mortality (Teriokhin et al., 2004), which would put evolutionary pressure on humans to reproduce at an earlier age (Stearns, 2000). Puberty itself does not predict AAO (Cohen, Seeman, Gotowiec, & Kopala, 1999; Ruiz, Blanco, Santander, & Miranda, 2000), but AAO might be related to sexual maturation of the brain and consequent mating, which might follow puberty by a shorter interval closer to the equator. For example, the interval between median age at menarche and median age at first live birth in 11 countries ranged from five years in Kenya to 11 years in Australia (Morabia & Costanza, 1998). We found a significant correlation between this interval and the latitude of the capital cities of the 11 countries (Pearson r = 0.72, t = 3.08, df = 9, p = .013), with shorter intervals closer to the equator. The gradient cannot be explained by greater use of oral contraceptives at higher latitudes, because the gradient was also present among women born before the advent of oral contraceptives. Thus, it is possible that greater pathogen diversity closer to the equator leads to earlier onset of schizophrenia by increasing mortality and thereby selecting for an earlier age reproduction. Greater pathogen diversity also might decrease the evolved age at reproduction through its effect on sexual selection and mating systems. Among non-human animals 6 females switch from monogamy to polygyny when males vary so much in quality (e.g., because of pathogen stress) that it is advantageous to mate with the highest quality male, even if that male must be shared with other females (Davies, 1989; Hamilton & Zuk, 1982). The same mechanism might underlie human polygyny, since monogamy is rare in areas of high pathogen stress, and pathogen stress explains 28% of the variation in degree of polygyny worldwide (Low, 1988). Because polygyny is associated with earlier sexual maturation in mammals, including humans (Kanazawa, 2001), this could lead to earlier reproduction and therefore earlier AAO. 4.3 Sexual selection for condition-sensitive fitness indicators Sexual selection also produces extravagant traits (e.g., attractive peacock tails) that function as fitness indicators—they vary greatly in some quality (e.g., size, loudness or complexity), and that variation correlates with underlying fitness (Pomiankowski & Moller, 1995; Rowe & Houle, 1996). The opposite sex evolves a preference for the highfitness, extravagant extreme (Kokko, Brooks, Jennions Michael, & Morley, 2003). We have proposed that schizophrenia is the low-fitness extreme of a sexually selected fitness indicator that evolved in humans through mutual mate choice (Shaner, Miller, & Mintz, 2004). According to our hypothesis, human embryos contain genetic instructions to grow brain systems specialized for particular forms of courtship; e.g., verbal courtship. Because these courtship behaviors evolved as fitness indicators, many fitness-reducing mutations and environmental hazards can disrupt development of the necessary brain systems and thereby reduce the attractiveness of courtship behavior. The result is large variation in the trait (e.g., verbal courtship skill) across individuals—variation that correlates with underlying fitness. Our hypothesis is that severe disruption of the necessary brain systems results in schizophrenia rather than in behaviors recognizable as courtship. If correct, our hypothesis could explain the latitudinal gradient in AAO, along with many seemingly-unrelated facts about the disorder. Its explanatory power arises from the generic properties of all sexually selected fitness indicators, two of which may help explain the latitudinal gradient in AAO. First, animals display fitness indicators most prominently after sexual maturation and during courtship and sexual competition. If schizophrenia is the low-fitness extreme of such an indicator, it should not be apparent to prospective mates until the age at which courtship and sexual competition usually begin. A latitudinal gradient in the age at onset of courtship and sexual competition thus would be expected to produce a similar gradient in AAO. Several proxies for initial sexual behavior, such as age at first sexual intercourse, are proximate determinates of fertility, as well as risk factors for other diseases, such as acquired immune deficiency syndrome. However, we have been unable to find such data, gathered by similar methods across a sufficient range of latitude, to test this prediction. Second, fitness indicators reflect individual fitness through their sensitivity to general phenotypic condition. We would predict a latitudinal gradient in AAO if there is a latitudinal gradient in general condition; e.g., due to differences in food intake or pathogen prevalence. Greater pathogen diversity towards the equator could increase the rate of infections, reduce overall phenotypic condition, and contribute to the latitudinal gradient in human mortality (Teriokhin et al., 2004). Because the indicator trait (e.g., verbal courtship skill) is expected to be condition-sensitive, this would increase the risk of schizophrenia and decrease AAO. This mechanism is much broader than the 7 one discussed above, in which prenatal infections predispose to schizophrenia. This is because fitness indicators should evolve to be sensitive not only to prenatal infections, but also to a large proportion of all the interactions between genes and environment that determine overall phenotypic condition. The incidence of schizophrenia is available for only eight of the cities in figure 1 (Bresnahan, Menezes, Varma, & Susser, 2003) and does not show a latitudinal gradient across these cities. However, none of the eight cities lies south of 21 degrees north latitude. The incidence of narrowly defined schizophrenia, determined by similar methods, is available for three Caribbean countries (Bresnahan et al., 2003) between 10 and 18 degrees north latitude. When these three countries are included, incidence of schizophrenia does correlate significantly negatively with latitude (Spearman r = -0.61, df = 9, p = .046). Third, fitness indicators should evolve greater fitness sensitivity especially under conditions of more intense sexual competition, for example in populations with high rates of polygyny and/or extra-pair copulations (“infidelity”) (Hasson, 1989; Pomiankowski & Moller, 1995; Rowe & Houle, 1996). This may explain why several studies have reported more favorable outcomes for schizophrenia in developing countries (Craig, Siegel, Hopper, Lin, & Sartorius, 1997; Sartorius et al., 1986). Developing countries are closer to the equator, so may have higher rates of infection, mortality and polygyny. If so, they would also have higher frequencies of genes that increase the fitness sensitivity of the indicator trait underlying schizophrenia. As a result, milder impairments in fitness and condition would produce the symptoms of schizophrenia, thereby elevating incidence while improving outcomes. 4.4 A test of competing causal hypotheses Sensitivity to underlying fitness and general condition also leads to a prediction unique to our hypothesis—the offspring of those who migrate in any direction (except back to ancestral homelands) will develop schizophrenia at a higher rate and with a lower AAO than offspring of those who do not migrate (Table 1). This is because individuals are likely to be best adapted to the environment of their ancestors. Migration to new environments exposes offspring to new hazards (e.g., pathogens, toxins, climates, and social exclusion) to which they are not well adapted. The resulting decrement in fitness and condition could disrupt the development of the indicator trait, increase the risk of schizophrenia, and decrease AAO as well. <Insert Table 1 about here> In contrast, if the environmental factor acts directly on individuals to influence AAO, then current latitude, not ancestral latitude, will determine AAO. Offspring of migrants to higher latitudes will develop schizophrenia later than offspring of those who remain at lower latitudes, and AAO of migrants will be typical of the higher latitude. Finally, if the environmental factor acts on evolution to vary the frequencies of genes that influence AAO but does not lead to the evolution of a fitness indicator, then genetic ancestry, not current latitude, will determine AAO. For example, if the environmental factor caused humans at lower latitudes to evolve to reproduce earlier, then offspring of migrants to higher latitudes will share these genes and develop schizophrenia at an age typical of that at lower latitudes. 8 Thus, in the case of migration towards the poles, the three mechanisms—direct environmental effects, natural selection for reproductive maturation rate, and sexual selection for condition-sensitive fitness indicators—make different and testable predictions about AAO in offspring. Only sexual selection for condition-sensitive indicators predicts earlier AAO. This is important, because several studies have shown a puzzling, several-fold increase in the rate of schizophrenia among the offspring of AfroCaribbean immigrants to European countries (Cantor-Graae & Selten, 2005). These reports do not contain data on AAO, but given the general association between risk and AAO, it is likely that AAO was earlier than expected. If so, then neither direct environmental effects nor evolved maturation rates can explain it. However, sexual section for fitness indicators could. 4.5 Life-history traits, fitness indicators and latitude The age distribution of reproductive effort and the degree of polygyny are among several “life-history traits” that evolve in response to a wide range of environmental factors (e.g., temperature, rainfall, photo-period, food supply, pathogens, and predators (Stearns, 1976). Similarly, fitness indicators reveal the extent to which individuals are well adapted to the environment in which they developed—the extent to which they have thrived despite environmental hazards. Thus, if schizophrenia is linked to “life-history traits” or fitness indicators, then its incidence and AAO will correlate with a wide range of environmental factors that can affect health and life span, many of which vary with latitude. We have considered pathogens and mortality, but the same may be true of others including poverty, malnutrition and murder. Conclusions We have discovered a potential clue to schizophrenia’s cause—its age at onset (AAO) varies linearly with latitude. People living close to the equator develop the illness about 10 years earlier than those living far from the equator. This is a large effect and understanding it could help explain the causes of schizophrenia. However, there are many potential causes because (1) the key environmental factor could be a feature of the physical environment (e.g., photo period, climate), the biological environment (e.g., pathogen diversity), or the social environment (e.g., degree of polygyny); (2) that factor could act directly on individuals, or indirectly on the evolution of traits (and genes) related to AAO; and (3) the factor could increase disease risk or accelerate the maturation of a trait related to schizophrenia. To illustrate several general mechanisms, we explored how one potential specific factor, pathogen diversity, might affect AAO. Those general mechanisms can be distinguished because they make different predictions about AAO among the offspring of those who migrate towards the poles, so this should be an initial focus of further research. In addition, two mechanisms predict a latitudinal gradient in the age at onset of reproductive behavior, and these could be tested using data on a proxy variable, such as age at first sex, gathered by similar methods across a broad range of latitude. 9 References Bresnahan, M., Menezes, P., Varma, V., & Susser, E. (2003). Geographical variation in incidence, course and outcome of schizophrenia: a comparison of developing and developed countries. In R. M. Murray, P. B. Jones, E. Susser, J. van Os & M. Cannon (Eds.), The Epidemiology of Schizophrenia (pp. 18-33). Cambridge: Cambridge University Press. Brown, A. S., Schaefer, C. A., Quesenberry, C. P., Jr., Liu, L., Babulas, V. P., & Susser, E. S. (2005). Maternal exposure to toxoplasmosis and risk of schizophrenia in adult offspring. Am J Psychiatry, 162(4), 767-773. Cantor-Graae, E., & Selten, J. P. (2005). Schizophrenia and migration: a meta-analysis and review. Am J Psychiatry, 162(1), 12-24. Cardno, A. G., Marshall, E. J., Coid, B., Macdonald, A. M., Ribchester, T. R., Davies, N. J., et al. (1999). Heritability estimates for psychotic disorders: the Maudsley twin psychosis series. Arch Gen Psychiatry, 56(2), 162-168. Carpenter, W., Jr. (2005). Unmet therapeutic needs. Schizophr Bull, 31(4), 793-794. Childs, B., & Scriver, C. R. (1986). Age at onset and causes of disease. Perspect Biol Med, 29(3 Pt 1), 437-460. Cohen, R. Z., Seeman, M. V., Gotowiec, A., & Kopala, L. (1999). Earlier puberty as a predictor of later onset of schizophrenia in women. Am J Psychiatry, 156(7), 1059-1064. Craig, T. J., Siegel, C., Hopper, K., Lin, S., & Sartorius, N. (1997). Outcome in schizophrenia and related disorders compared between developing and developed countries. A recursive partitioning re-analysis of the WHO DOSMD data. Br J Psychiatry, 170, 229-233. Davies, N. B. (1989). Sexual Conflict and the Polygamy Threshold. Animal Behaviour, 38(2), 226-234. DeLisi, L. E. (1992). The significance of age of onset for schizophrenia. Schizophr Bull, 18(2), 209-215. Dubertret, C., Gorwood, P., Gouya, L., Deybach, J. C., & Ades, J. (2001). Association and excess of transmission of a DRD2 haplotype in a sample of French schizophrenic patients. Schizophr Res, 49(1-2), 203-212. Gangestad, S., & Buss, D. M. (1993). Pathogen prevalence and human mate preferences. Ethology and Sociobiology, 14(2), 89-96. Guernier, V., Hochberg, M. E., & Guegan, J. F. (2004). Ecology drives the worldwide distribution of human diseases. PLoS Biol, 2(6), e141. Hamilton, W. D., & Zuk, M. (1982). Heritable true fitness and bright birds: a role for parasites. Science, 218, 384-387. Hasson, O. (1989). Amplifiers and the Handicap Principle in Sexual Selection a Different Emphasis. Proc R Soc Lond B Biol Sci, 235(1281), 383-406. Haverkamp, F., Propping, P., & Hilger, T. (1982). Is there an increase of reproductive rates in schizophrenics? I. Critical review of the literature. Archiv fuer Psychiatrie und Nervenkrankheiten, 232(5), 439-450. Jablensky, A. (2000). Epidemiology of schizophrenia: the global burden of disease and disability. Eur Arch Psychiatry Clin Neurosci, 250(6), 274-285. Jablensky, A., & Cole, S. W. (1997). Is the earlier age at onset of schizophrenia in males a confounded finding? Results from a cross-cultural investigation. Br J Psychiatry, 170, 234-240. 10 Jablensky, A., Sartorius, N., Ernberg, G., Anker, M., Korten, A., Cooper, J. E., et al. (1992). Schizophrenia: manifestations, incidence and course in different cultures. A World Health Organization ten-country study. Psychol Med Monogr Suppl, 20, 1-97. Kanazawa, S. (2001). Why father absence might precipitate early menarche: The role of polygyny. Evolution & Human Behavior, 22(5), 329-334. Kokko, H., Brooks, R., Jennions Michael, D., & Morley, J. (2003). The evolution of mate choice and mating biases. Proc R Soc Lond B Biol Sci, 270(1515), 653-664. Low, B. S. (1988). Pathogen Stress and Polygyny in Humans. In L. Betzig, M. B. Mulder & P. Turke (Eds.), Human Reproductive Behaviour: A Darwinian Perspective. Vii+363p. , Uk. Illus. Maps (pp. 115-128). Cambridge: Cambridge University Press. Marcelis, M., Navarro-Mateu, F., Murray, R., Selten, J. P., & van Os, J. (1998). Urbanization and psychosis: A study of 1942-1978 birth cohort in the Netherlands. Psychological Medicine, 28(4), 871-879. McDonald, C., & Murphy, K. C. (2003). The new genetics of schizophrenia. Psychiatr Clin North Am, 26(1), 41-63. Morabia, A., & Costanza, M. C. (1998). International variability in ages at menarche, first livebirth, and menopause. World Health Organization Collaborative Study of Neoplasia and Steroid Contraceptives. Am J Epidemiol, 148(12), 1195-1205. Nunn, C. L., Altizer, S. M., Sechrest, W., & Cunningham, A. A. (2005). Latitudinal gradients of parasite species richness in primates. Diversity and Distributions, 11(3), 249-256. Pomiankowski, A., & Moller, A. P. (1995). A resolution of the lek paradox. Proc R Soc Lond B Biol Sci, 260(1357), 21-29. Rowe, L., & Houle, D. (1996). The lek paradox and the capture of genetic variance by condition dependent traits. Proc R Soc Lond B Biol Sci, 263(1375), 1415-1421. Ruiz, A., Blanco, R., Santander, J., & Miranda, E. (2000). Relationship between sex differences in onset of schizophrenia and puberty. Journal of Psychiatric Research, 34(4-5), 349-353. Sartorius, N., Jablensky, A., Korten, A., Ernberg, G., Anker, M., Cooper, J. E., et al. (1986). Early manifestations and first-contact incidence of schizophrenia in different cultures. A preliminary report on the initial evaluation phase of the WHO Collaborative Study on determinants of outcome of severe mental disorders. Psychol Med, 16(4), 909-928. Shaner, A., Miller, G., & Mintz, J. (2004). Schizophrenia as one extreme of a sexually selected fitness indicator. Schizophr Res, 70(1), 101-109. Stearns, S. C. (2000). Life history evolution: Successes, limitations, and prospects. Naturwissenschaften, 87(11), 476-486. Stevens, J. R. (2002). Schizophrenia: reproductive hormones and the brain. Am J Psychiatry, 159(5), 713-719. Teriokhin, A. T., Budilova, E. V., Thomas, F., & Guegan, J. F. (2004). Worldwide variation in life-span sexual dimorphism and sex-specific environmental mortality rates. Hum Biol, 76(4), 623-641. Thaker, G. K., & Carpenter, W. T., Jr. (2001). Advances in schizophrenia. Nat Med, 7(6), 667-671. Tochigi, M., Okazaki, Y., Kato, N., & Sasaki, T. (2004). What causes seasonality of birth in schizophrenia? Neuroscience Research, 48(1), 1-11. 11 Verdoux, H., Geddes, J. R., Takei, N., Lawrie, S. M., Bovet, P., Eagles, J. M., et al. (1997). Obstetric complications and age at onset in schizophrenia: an international collaborative meta-analysis of individual patient data. Am J Psychiatry, 154(9), 1220-1227. The World Factbook. (Accessed March, 2006). http://www.cia.gov/cia/publications/factbook/index.html. 12 Figure 1: Latitude and age at onset of schizophrenia 36 Age at onset (years) 34 32 30 28 26 24 22 0 10 20 30 40 50 Latitude of study city (degrees) 60 13 Fig. 1. Latitude and Age at Onset of Schizophrenia in 13 Northern Hemisphere Cities. Data drawn from (Jablensky & Cole, 1997). In order of increasing latitude the cities were Cali, Colombia; Ibadan, Nigeria; Agra, India; Honolulu, U.S.A.; Chandigarh, India (rural and urban); Nagasaki, Japan; Rochester, U.S.A.; Prague, Czech Rep.; Nottingham, U.K.; Dublin, Ireland; Moscow, Russia and Aarhus, Denmark. Bars represent ± 1 s.e.m. 14 Table 1: Predicted change in AAO among offspring of migrants* Mechanism responsible for latitudinal gradient Direction of migration Equatori East or Polar al west Direct environmental none effect on individuals Natural selection for reproductive none none none maturation rate Sexual selection for condition-sensitive fitness indicators *Change indicated as increased (), decreased () or unchanged (none) compared with AAO among offspring of those who did not migrate.